Deletion of the RNA-binding proteins ZFP36L1 and ZFP36L2 leads to perturbed thymic development and T lymphoblastic leukemia

Abstract

ZFP36L1 and ZFP36L2 are RNA-binding proteins (RBPs) that interact with AU-rich elements in the 3' untranslated region of mRNA, which leads to mRNA degradation and translational repression. Here we show that mice that lacked ZFP36L1 and ZFP36L2 during thymopoiesis developed a T cell acute lymphoblastic leukemia (T-ALL) dependent on the oncogenic transcription factor Notch1. Before the onset of T-ALL, thymic development was perturbed, with accumulation of cells that had passed through the beta-selection checkpoint without first expressing the T cell antigen receptor beta-chain (TCRbeta). Notch1 expression was higher in untransformed thymocytes in the absence of ZFP36L1 and ZFP36L2. Both RBPs interacted with evolutionarily conserved AU-rich elements in the 3' untranslated region of Notch1 and suppressed its expression. Our data establish a role for ZFP36L1 and ZFP36L2 during thymocyte development and in the prevention of malignant transformation.

Figure 1. Zfp36l1-Zfp36l2 dKO mice develop T-ALL

A. Kaplan Meier Survival of dKO (red line, n=37), Zfp36l2 single knockout (dashed line, n=14) and control (blue line, n=30) mice. B. Flow cytometric analysis of control thymus and a dKO thymic tumor stained with indicated antibodies. sTCR-β = surface TCR-β, icTCR-β = intracellular TCR-β. Solid red line = icTCR-β, broken blue line = isotype control. C. Low power examination of peripheral blood from diseased dKO mouse showing leukocytosis. D. High power examination of peripheral blood from diseased dKO mouse showing circulating lymphoblasts. E. Flow cytometric analysis of thymus, spleen and marrow from a representative control and a diseased dKO mouse. F. Dβ2-Jβ2 rearrangement by PCR of DNA from control kidney, three control thymi, five dKO thymi (12-weeks old) and four dKO thymic tumors.

Figure 2. Thymocyte development is perturbed prior to tumor development

A. Flow cytometric plots gated on cells negative for dump channel (B220, Ter119, NK1.1, γδ-TCR, Mac-1, Gr1). B. Total thymic cellularity showing thymic atrophy in dKO mice. Statistical analysis used Mann-Whitney test. C. Flow cytometric plots gated on CD8 single positive population (CD8⁺CD4⁻ as shown in A) showing proportions of immature (CD24^hisTCRβ⁻) and mature (CD24^intsTCRβ⁺) CD8 single positive thymocytes. D. Immature CD8 single positive thymocytes as a percentage of all CD8 SP cells (upper) or absolute number per thymus (lower) at the indicated age of mice. Graphs show the mean and SEM for five mice per genotype. E. Flow cytometric plots from 3-week old mice gated on dump channel negative, CD4⁻CD8⁻ double negative thymocytes. For all flow cytometric plots numbers show the percentage of cells within the indicated gate and represent the mean from one experiment of 4–5 mice per genotype. All plots are representative of at least nine individual mice.

Figure 3. Expression of Notch1 is elevated in Zfp36l1-Zfp36l2 dKO mice

A. Heatmap summary of microarray performed on cDNA from whole thymus of control and dKO mice at five and nine weeks of age. Criteria for inclusion are genes changing significantly (p<0.05), elevated > 1.5 fold in dKO mice at both five weeks and nine weeks relative to control thymus. B. Real-time PCR performed upon cDNA from whole thymus for Notch1 and Notch target genes. Values are normalised to expression of B2m and presented relative to expression of each gene in thymocytes from control mice. Graphs show mean and SEM of 3–8 mice. C. Immunoblot for the cleaved intracellular Notch1 (iCN1) performed on whole thymus from control or diseased double knockout mice. D. Notch1 mean fluorescence intensity (MFI) by flow cytometry (normalised to isotype control) in control thymus and dKO thymic tumors. E. Notch1 expression in double negative, double positive, CD8 and CD4 populations using the gating strategy shown. Mean and SEM are shown for five separate thymic tumors. F. Notch1 expression by flow cytometry on thymocyte populations gated as shown in 3-week old. Upper plots are gated on dump channel (B220, Ter119, NK1.1, γδ-TCR, Mac-1, Gr1) negative. Lower plots gated on dump⁻CD4⁻CD8⁻ DN thymocytes and show the gating of populations A, B & C. G. Notch1 expression by flow cytometry in CD8 subpopulations (D, E & F) from 3-week old mice using gating strategy indicated. Unless stated, error bars show the mean and standard error of four-five mice per genotype. Statistical analysis performed using Mann Whitney test. (* P < 0.05, ** P < 0.01)

Figure 4. ZFP36L1 and ZFP36L2 exert suppression via interaction with sequences in the Notch1 3'UTR

A. Notch1 3'UTR presented as degree of conservation between human and mouse using the Vista Genome Browser. The highly conserved region (HCR) of interest is underlined in red. B. Detailed view showing inter-species sequence conservation for the underlined region in A. Black background indicates 100% conservation between human, mouse, dog, armadillo, opossum and platypus. Predicted ZFP36L1 and ZFP36L2 binding sites are boxed in red. C. Luciferase reporters were constructed corresponding to the full length Notch1 3'UTR (N1 UTR full length), the proximal HCR underlined in red in A (N1 HCR), or the Notch1 3'UTR with the HCR removed (N1 Δ HCR). These reporters were co-transfected into HEK293T cells along with a pCDNA3 empty control vector or pCDNA3 expressing ZFP36L1, ZFP36L2 or a tandem zinc finger mutant of ZFP36L1 (TZFM). Results are shown as the mean and SEM of five separate transfections and are representative of three experiments. Statistical analysis performed by ANOVA with Tukey post-hoc analysis (* P < 0.001) D. Titration of transfected ZFP36L1 or ZFP36L2. Mean and SEM of five separate transfections are shown. E. Electromobility shift assay (EMSA) following incubation of a radiolabelled Notch1 probe (corresponding to the 61 nucleotides containing the nonameric AU sequence) with lysates from 293T cells transfected with the indicated expression constructs. pCDNA3 = empty control vector, TZFM = tandem zinc finger mutant.

Figure 5. Inhibition of Notch or re-expression of ZFP36L1 is toxic to tumor growth

A. Primary dKO tumor cells were cultured on OP9-DL1 stromal cells for three days in the presence of increasing concentrations of the gamma-secretase inhibitor Compound E. Graph shows mean and SEM for six tumors presented relative to the number of cells seeded. Statistical analysis was performed by repeated measures ANOVA with Tukey post hoc analysis. (**p<0.01, ***p<0.001) B. Notch1 blocking antibody, or saline control, was administered in vivo to 11 week old control and dKO mice. After three weeks of treatment thymus was analysed by flow cytometry. Post-treatment thymic cellularity is shown for individual mice. C. Representative flow cytometry plots of thymus after treatment with control or Notch1 blocking antibody. CD4–8 plots are gated on thymocytes negative for B220, NK1.1, γδ-TCR, mac-1, Gr1, Ter119. CD24-icTCR-β plots are gated on the CD8 single positive quadrant. D. Primary thymic tumors from dKO mice were cultured on OP9-DL1 stromal cells for 24 hours before infection with equivalent titres of retrovirus expressing either ZFP36L1 or the tandem zinc finger mutant of ZFP36L1 (TZFM). After three days cells were analysed by flow cytometry. Flow cytometry plots shown are representative of infection of five separate dKO tumors.